Re-Ranking High-Dimensional Deep Local Representation for NIR-VIS Face Recognition
Re-Ranking High-Dimensional Deep Local Representation for NIR-VIS Face Recognition
复制标题
重新排序 NIR-VIS 人脸识别的高维深度局部表示
DOI：
10.1109/tip.2019.2912360
复制
发表时间：
2019-09-01
影响因子：
10.6
通讯作者：
Gao, Xinbo
中科院分区：
文献类型：
--
作者：
Peng, Chunlei;Wang, Nannan;Gao, Xinbo
异构人脸识别是指对从不同传感器或来源捕获的人脸图像进行匹配，在公共安全和执法领域有着广泛的应用。由于感知和创建过程的巨大差异，异构面部图像之间存在巨大的特征差距。现有的方法仅仅关注在特征空间中将探测图像与图库进行比较，而由于不同的传感模式引起的外观变化，真正的目标可能不会出现在第一排。为了利用初始排序结果中的有价值的信息，本文提出对高维深度局部表示进行重新排序，以匹配近红外（NIR）和视觉（VIS）面部图像，即NIR-VIS人脸识别。首先通过卷积神经网络（CNN）提取和连接局部面部斑块上的深层特征来构建高维深层局部表示。通过比较压缩的深度特征可以得到初始的NIR-VIS识别排名结果。然后，我们提出了一种新颖且高效的局部线性重排序（LLRe-Rank）技术来细化初始排序结果，该技术可以从初始排序结果中探索有价值的信息。所提出的重新排序方法不需要任何人工交互或数据注释，并且可以用作无监督的后处理技术。在最具挑战性的Oulu-CASIA NIR-VIS数据库和CASIA NIR-VIS 2.0数据库上的实验结果证明了我们方法的有效性。
Heterogeneous face recognition refers to matching facial images captured from different sensors or sources, which has wide applications in public security and law enforcement. Because of the great differences in sensing and creating procedure, there is a huge feature gap between heterogeneous facial images. The existing methods merely focus on comparing the probe image with the gallery in feature space, while the true target may not appear at the first rank due to the appearance variations caused by different sensing patterns. In order to exploit valuable information from the initial ranking result, this paper proposes to re-rank high-dimensional deep local representation for matching near-infrared (NIR) and visual (VIS) facial images, i.e., NIR-VIS face recognition. A high-dimensional deep local representation is first constructed by extracting and concatenating deep features on local facial patches via a convolutional neural network (CNN). The initial NIR-VIS recognition ranking results can be obtained by comparing the compressed deep features. We then propose a novel and efficient locally linear re-ranking (LLRe-Rank) technique to refine the initial ranking results, which can explore valuable information from the initial ranking result. The proposed re-ranking method does not require any human interaction or data annotation and can be served as an unsupervised postprocessing technique. The experimental results on the most challenging Oulu-CASIA NIR-VIS database and CASIA NIR-VIS 2.0 database demonstrate the effectiveness of our method.